Just as meteorologists analyse atmospheric data patterns to predict future weather conditions, diagnostic technologies pick up distinct motifs in patient biomarkers to detect disease. In the case of lung cancer diagnostics, Raman spectroscopy has emerged as a promising method of analysing subtle changes in the biochemistry of the fluid surrounding the lungs.
Raman spectroscopy uses laser light to interact with a blood or fluid sample, and by analysing the unique shifts in frequency of the scattered light—known as Raman scattering—creates a molecular ‘fingerprint’ that can be analysed using computational methods. However, in practice, traditional methods for diagnosing lung cancer from Raman spectra can be subjective, time-consuming and inaccurate.
Malini Olivo, who leads the Translational Biophotonics Laboratory at A*STAR Skin Research Labs (A*SRL) where she is a Distinguished Principal Scientist, said that label-free surface-enhanced Raman spectroscopy, or SERS, offers unique advantages for catching lung cancer early.
“Unlike traditional methods that rely on labelling with specific markers which has the possibility [of] poor specificity due to non-specific binding of the labels, label-free SERS directly detects molecular vibrations of specific biomolecules, providing a more rapid and sensitive approach,” explained Olivo, adding that the technique is also non-invasive, sparing patients from painful biopsies.
In collaboration with researchers from the National University Hospital, Singapore, Olivo and A*SRL Senior Scientist, Jayakumar Perumal, tested a novel approach of using SERS as a means of improving the sensitivity of conventional Raman spectroscopy for more sensitive and accurate lung cancer diagnoses.
They collected pleural effusion samples from healthy controls and patients with different medical conditions including lung and breast cancer. Over 500 SERS spectra were acquired for each sample and machine learning techniques were used to classify the samples based on their Raman spectra signatures.
In the study, the team reported an impressive 85 percent classification accuracy for distinguishing lung cancer cases using their new approach. The technique also had an 87 percent sensitivity for correctly identifying lung cancer and an 83 percent specificity for classifying healthy controls.
Olivo said that these findings suggest that SERS may become a new gold standard for lung cancer and beyond. “Since different cancers have distinct molecular profiles, the methodology's ability to detect specific Raman spectral patterns in biofluids could potentially enable the development of diagnostic algorithms for classifying different cancer subtypes.”
The team’s current efforts are focused on validating their findings in a larger patient cohort as well as adapting the methodology for infectious disease diagnostic applications.
The A*STAR-affiliated researchers contributing to this research are from A*STAR Skin Research Labs (A*SRL).
